‍

Most hoteliers want guests to book through direct channels rather than through third-parties such as Online Travel Agencies (OTAs). Loyalty programs, member-only rates, and search-engine advertising have commanded increasing shares of the marketing budget of many hotels trying to achieve this objective.

But surprisingly few hotels have focused on the blocking and tackling of guest engagement with direct channels, and that will be my topic this week. I refer to techniques that:

• Increase the conversion rate by keeping visitors on your website or in your booking engine rather than looking elsewhere

• Tailor product offerings to better appeal to each visitor

• Focus visitors on higher-value offerings as opposed to lowest price, and

• Bring them back to your website after they leave.

‍

I will also include some tools that enable hotel reservations offices to automate routine operations so that agents can spend more time on sale opportunities where they can make a difference. Some of these can also improve conversion rate by enabling agents to respond to written requests much faster.

Much has been written elsewhere about website chatbots, and my last blog addressed opportunities with artificial-intelligence (AI) powered contact center technologies. However, this week I will focus on several other solutions that hotels and resorts are using successfully, but that are not yet mainstream. Some of these are simple and relatively inexpensive, while others are powerful and may make financial sense mostly for certain types of properties. Most of the capabilities can be incorporated into existing booking engines and (in some cases) other reservations software such as chatbots or back-office reservations operations.

Conversion rate and the direct-booking ratio are tightly intertwined. When your website or mobile app has the chance to make a sale but fails to do so, then the customer either chose a different hotel, or booked your hotel but on a different site.  You may or may not have had any real chance to convert the customer who ended up choosing a different hotel (that depends on the reason they chose it), but abandonment of the hotel booking site for a third-party channel is often preventable.

There are hundreds of hotel booking engines in the market, a few of which provide some of the capabilities I will cover here. But most do not, or offer only limited subsets. I have therefore focused on companies whose capabilities can be plugged into any of the common booking engines. This means they will be viable options for most hotels and groups. And because products like this tend to be laser-focused on doing a few things really well, they are therefore more likely to have developed and incorporated best practices. If your current solution includes some of these capabilities, you can use this material to help evaluate its feature set.

Senior executives from several innovative companies were instrumental to my research for this article. I want to thank those from Cartstack, Hotel Res Bot, Hotelverse, and 123compare.me.

I also got some great material from P3, which does not offer any tools independently of their own booking engine, but which was the lone customer I could find of some interesting technology previously offered by Arvoia that is no longer available on the market. The lessons were still very useful; P3’s conversion and upselling success using even just the earliest capabilities of Arvoia’s technology has left them wanting more (and unable to find it). The technology worked even if the company did not. Perhaps someone will fill that gap?

All but one of the companies included this week are based in Europe. The more fragmented booking landscape in Europe creates more opportunities for innovative products than many other markets, and especially North America. Europe’s proliferation of independent hotels has encouraged vendors to try new approaches that may be rejected as too costly or disruptive by major brands that would need to shoehorn them into legacy systems. That is a barrier to the commercial success of products targeting the brand-dominated North American market.

For easier reading, I have divided the tools into categories based on the specific issues they address, but there is overlap and any such division is going to be imperfect. I have mentioned vendors who showed me or discussed specific capabilities, but that does not mean that other products (whether mentioned here or not) do not offer similar capabilities.

Reengagement and Cart Recovery

Tools in this category come into play when users of a website and/or web booking engine leave without completing a booking.  This may occur before starting a booking (referred to as browse abandonment) or after starting one but before completing it (booking abandonment). In the latter case, details such as dates, the size and makeup of the party (e.g. adults and children), and certain user characteristics (such as country, browser type, or device type) are typically available. Depending how far the user got in the booking process, they may have also provided other information such as name, email address, or phone number.

For browse abandonment, the techniques are somewhat limited. If the user leaves your site by opening a new tab (perhaps to look at another hotel or booking channel, or maybe because they got distracted by another task), tools like Cartstack can provide reminders as long as the user does not close the tab for your site. They can, for example, change the icon or name on the tab itself, and/or sound a chime. And if a user closes the tab without booking, they can pop up a window asking if the user wants to continue later (and collect their email address if they do). Then they can send an email with a link that can effectively re-establish their session where they left off.

Alternatively, popup windows can be generated when a user tries to exit, potentially offering a better deal to stimulate conversions; both Cartstack and 123compare.me supported this. The offer may be targeted at specific time periods, products, or customer profiles, or they may offer services that reduce the perceived price or risk (such as buy-now-pay-later, weather guarantees for a resort, or trip insurance).

Exit popups may also provide information such as parity rates or review summaries from several online sites, to help nudge the user to complete the booking rather than doing further research on other sites. Images and text may be personalized; for example, a family visiting a beach resort might see visuals of childrens’ activities, while a weekend couple gets the pictures of the romantic restaurant.

Attempts to recover booking abandonment can be more aggressive than for browser abandonment, and several of the companies I spoke with used similar techniques. Integration with the web booking engine (often quite simple) will enable them to capture whatever information the user has entered, even before the first form has been completed and submitted. I was told that once a user starts a booking, there is a roughly even chance that they will enter an email address prior to abandonment (and the probability can be maximized if the email is one of the first fields requested). Hotel Res Bot did not support this for web booking engines, but did offer a similar capability for chatbot conversations that did not result in bookings.

Even if the user does not get far enough to enter an email address, contact information may be available if they are logged in, or if they have provided it on a prior visit to the site and have not cleared cookies since.

For abandoned bookings, the tools will typically gather information entered on the booking engine related to the stay, such as the specific hotel, the dates, the size and composition of the party, the number of rooms, the room type, and the rate. The tools will then send an email to the user, which will minimally provide a link back to a prepopulated page in the booking engine where they can complete the booking. Some tools allow the hotel to sweeten the offer, based on demand, on characteristics of the booking party, or other factors. The follow-up email is triggered when the tool detects that a user started a booking but never reached the final confirmation page.

Vendors report high conversion rates on these efforts, although it is difficult to determine how many bookers would have returned to the site anyway, so not all of those conversions may be incremental. But common sense suggests that some will be.

Site Stickiness

The tools offered different capabilities to keep guests engaged on the site; the longer a guest spends on a site, the better the conversion (and also the better the site will be positioned by major search engines).

For resorts and other complex hotels, immersive digital experiences were key. Hotelverse, for example, produces digital twins that enable users to fly over the entire property, look at specific buildings or even rooms, and drill down to images and videos of key features such as pools or restaurants. You can see an example here; scroll down to the “Guarantee yourself the perfect room” section, enter some dates, activate the image, and click wherever you see an icon or a price.  If a user reaches the web booking engine first, then they see a dropdown option labeled “Reserve the room you want!”

Such approaches can be truly immersive, but they do require a lot of original content that either the vendor or the hotel will need to collect, and this can increase the cost. But for the right type of property (particularly large, sprawling resorts), properly implemented, conversion has reportedly improved by 30-60%, and time spent on site visits roughly doubles. That kind of result can easily justify the cost of collecting the content.

Stickiness is also increased by offering capabilities that users would otherwise have to leave the site to find. Three key widgets that do this are:

• Showing comparable rates from third-party booking sites to provide assurance that the user is getting the best rate. These should preferably be shopped in real time rather than by caching periodic updates. Some hotels may also want to adjust their own quoted rate if it is above any of the third-party quotes (123compare.me can do this).

• Showing a summary for ratings (or an AI-generated summary of actual reviews) from common review sites such as Google, Booking.com, and Tripadvisor. 123compare.me’s tool can filter and personalize ratings based on attributes of the user (for example, showing reviews for couples when the booking is for two adults over a weekend). Ideally the widget also supports drill-down to see individual reviews.

• Maps. We all know that the three most important factors in selecting a hotel are location, location, and location. Yet how many sites force the user to exit to Google Maps to find out what the hotel is near. Unlike a custom map shown from the hotel site, Google Maps will also show other hotels that might be more conveniently located, increasing the likelihood of abandonment.

‍

Personalization and Segmentation

Some booking engines, as well as third-party add-ons like 123compare.me, offer the ability to generate popups on the website or booking engine that can be used for targeted marketing campaigns. The hotel selects a template, modifies photos, fonts, text, size, and positioning, identifies a target audience (which may be based on the booking profile, need dates, or both), specifies when they should be displayed (for example, 20 seconds after the user lands on a page, or as they exit), and provides a call to action. As before, what a family on a week’s summer holiday sees might be totally different than a lone traveler on a weeknight in midwinter.

While this requires deeper booking engine integration, personalization and segmentation can also be applied to customize the selection and order of presentation of room types and rate plans. Expert opinions differ, but most agree that a small number (typically 5-10) of relevant choices is much better than dozens of irrelevant ones (if you really need more, you can include a “see more options” link at the bottom of the page).

The order matters as well; while it may be critical to have a parity offer matching OTA rates on the first screen, AI testing I have seen in various tools suggests that for most audiences, the first rate should be moderately higher than the parity, as that puts the user in a psychological state where higher rates are associated with higher value, not just higher cost.

Tools like Hotelverse take personalization to a new level by enabling the user to tour the hotel visually from the outside, pick the building that best meets their needs, see the rooms (and rates) for the different views based on side of the building or floor, view specific photos of rooms, and ultimately select the exact room that best meets their needs and preferences.

Arvoia is no longer around today, but was doing some very interesting things in this arena, with AI used to test various segmentation approaches and extensive A/B testing to optimize ordering. P3 told me that by putting a higher “anchor” rate first and the parity rate second or third, they increased the average booking value by 6%.

Hotel Res Bot uses a similar approach when creating a booking link to send in response to an email, web form, or chatbot booking request, filtering the offer based on both explicit and implicit requests by the guest as well as the hotel’s availability and selling strategy. They allow agents to review and approve the choice, but in most cases the agents accept the bot’s proposal.

Finally, personalization can be supported by attribute-based selling, by letting the customer select the specific room features they want. I covered that topic in detail about a year ago in a two-part series here and here, so I will not address that in detail today.

Operational Tools

Several tools help to automate or speed up repetitive manual tasks in reservations offices. These can both reduce the agent workload, help them convert more bookings, and free up agents to focus on higher value sales opportunities.

In some parts of the world (especially Europe, where small hotels dominate), manual reservation requests are still common. According to an exhaustive recent study authored by Prof. Roland Schegg, email accounts for 17% of European hotel bookings, web booking engines 11%, and contact-us forms on websites 6%. Tools like Hotel Res Bot can draft responses for emails and contact-us form requests and (depending on the specific situation) send them automatically if the hotel so wishes, or present them to an agent to approve or modify first.

Where bots can be trained to respond automatically for some use cases, they provide an additional conversion benefit in that they operate 24 hours a day, seven days a week, and can respond in a matter of seconds. Email and web requests that are not answered within a short period of time often lead to the guest booking a different hotel, or the same hotel but via a third-party channel. Speed matters.

At the same time, the bots are not experts; good agents are much better. Hotel Res Bot confirmed that in their experience (similar to hotels’ experience with chatbots), the best way to implement them is to make them a tool for the agents first, and then let them decide over time what the bots should be allowed to respond to autonomously. And if the bot’s response generates a nontrivial rejoinder from the user (like anything other than “thanks”), that response should almost always be routed to a person, as at least some portion of the time it will mean that the bot did not correctly understand or action the request.

Hotel Res Bot is also about to launch an automation tool for bill-to-company rates, which in Europe are very common and typically handled by an email request followed by a complex paper/scan/fax process today that can require a lot of tedious work (and paper files) at both the corporate travel department and the hotel reservation office. The product automates the response by sending a link that allows an authorized person at the company to approve the booking electronically.

When a tool creates a draft response for human review, integration with the tools that manage communications (such as Outlook, a Customer Relationship Management system, or a help-desk ticketing system) are important. Ideally the agent should be able to modify key fields (such as room type), with related fields (such as the associated rates or restrictions) automatically refreshed in the background. With the right integration, this can reduce the average response time from 2-3 minutes (or more) to a few seconds on many transactions.

Tools like this tend to be custom designed for hospitality rather than generic. Text requests for bookings (whether by email, contact-us form, or web chat) may have challenging combinations of key fields; for example, if a message contains four dates, do they represent two different stays for the same guest, different arrival and departure dates for guests who will have separate rooms, alternative dates in case one pair cannot be accommodated, or four different one-night stays where only the arrival dates were provided?

Additionally, such requests contain personally identifiable information, meaning that they cannot simply be fed into commercial AI tools such as ChatGPT without violating privacy regulations in many jurisdictions.

Conclusion

Hotels are most profitable when they can convert the most business, through the lowest-cost channels, with the least time and effort on the part of human agents. This week’s article has explored tools that support some or all of these objectives. If your hotel hasn’t looked at the category in the past few years, you should be asking yourself why not. They continue to get more powerful, and when applied properly to the right types of hotels, they can become cash machines.

‍